The Research On Texture Feature Extraction In Image Retrieval

The multimedia resources have appeared in a rash along with the rapid development of computer technology and the internet. It’s a hot question in multimedia technology research field that how to retrieval the interesting resources effectively. In image retrieval field, used the content based image retrieval technology to resolve some question about image retrieval. Image feature extraction technology is one of the critical technologies to struct image feature database. The texture feature data is one of the three major low level features of image. Thus, this topic is mainly to research the texture feature extraction technology of the image retrieval technology.First, this paper used the co-occurrence matrix based on texture primitive to extract texture feature of image. In this method, it extracted basic texture primitive of image by Local Binary Pattern (LBP), and used co-occurrence matrix of Gray Level Co-occurrence Matrix (GLCM) to analyze the texture primitive image. Then combined the thought of Gabor multi-resolution analysis and the method of primitive co-occurrence matrix, put forward the texture feature extraction method based on Gabor domain texture primitive co-occurrence matrix, and combined Gabor domain’s coefficients mean and variance, established the new texture feature vector.This paper put forward texture feature extraction algorithm based on Curvelet domain statistical model from below:changing the texture analysis from Curvelet and the coefficient distribution characteristics from Curvelet transformation. This algorithm transformed the question of texture feature extraction to the coefficient statistics model and the model parameters estimation.This paper analyzed experimentally the proposed algorithm by a standard Brodatz texture image database and the texture database contained20kinds of texture images, only to test the efficiency and robustness of the proposed algorithm. The experimental results showed that it has a good performance when the texture feature extraction was used to texture retrieval.